Research theme
The clearest statement of Felton’s central methodological argument, and the paper where the individual-specific position is stated most forcefully. The opening claim: because experimental fast-bowling research averages across bowlers, “the effect of individual technique parameters on ball speed has provided contradictory arguments” — a group study cannot tell you what this bowler should change. So: build a model of the individual, optimise it, and give that individual their own coaching prescription.
Method, plainly: a 16-segment planar torque-driven computer simulation model of the front foot contact phase built in AUTOLEV, with 14 rigid segments (head+trunk, two upper arms, two thighs, two shanks, two two-segment feet, forearm+hand on the non-bowling arm, forearm and hand on the bowling arm), wobbling masses in shanks/thighs/trunk, nine monoarticular joint torque generators (front MTP, front ankle, front knee, both hips, both shoulders, bowling elbow, bowling wrist), three ground-contact points per foot (heel, MTP, toe), and two massless segments allowing non-coincident hip and shoulder joint centres. Customised to one elite bowler; evaluated by matching a recorded delivery; then three optimisations.
The evaluation objective function was a six-component RMS difference: force, centre-of-mass velocity, orientation angle, ball speed, time, and the nine torque-driven joint angles — weighted equally, with 1° treated as equivalent to 1%. Penalties limited horizontal slide and vertical compression of the front foot at impact, wobbling-mass movement, and joint angles exceeding anatomical bounds.
What they measured
- How fast the ball leaves the hand (ball release speed).
- How bent the front knee is through the phase (front leg kinematics).
- How far the bowler folds forward (trunk flexion).
- When the bowling arm starts to come over (onset of bowling arm circumduction).
- When the front arm starts to come down (non-bowling shoulder extension timing).
- The pose at the instant the front foot lands (initial body configuration).
- How hard the front foot hits the ground (ground reaction force — in the objective function).
- How strong each joint is (subject-specific torque–angle and torque–angular-velocity relationships from a nine-parameter function).
Findings
All causal within the model, single bowler.
Model evaluation: 4% overall difference, with the kinematic components averaging 1%. The authors conclude the model can accurately reproduce the kinematics of the front foot contact phase and is fit for optimisation.
Optimising movement only (landing pose held at the bowler’s actual): +10%. Achieved “by maintaining a straighter front leg and increasing the amount of trunk flexion”.
Optimising the landing pose as well: +22%. “The most marked difference in the initial body configuration was at the shoulders where the extension was delayed for both the bowling and non-bowling arms. Adopting this initial body configuration allowed the front leg to stay straighter and more trunk flexion to occur.”
Increasing strength by 5% (ankle, knee, hip, shoulder): +1% relative to the optimal technique. The increase “allowed the individual to keep a straighter front leg, delay trunk flexion and produce more extension of the front arm”. The authors’ explanation for how small this is: the bowler is already inside an elite environment with strength and conditioning specifically designed to maximise fast bowling performance.
The explicit conclusion — technique, not strength. “The technique and strength optimisations indicated that the performance of the individual in this study is limited by his technique rather than strength. It is recommended that the future coaching of this individual is focussed on adapting his technique to keep his front leg straight, delay the bowling and non-bowling arms and increase trunk flexion whilst maintaining his current strength.”
Agreement with the group literature. The authors note the prescription matches Worthington et al. (2013), who argued that elite bowlers use a straight front leg to more efficiently convert the linear momentum of the run-up into angular momentum, which in turn produces increased trunk flexion and a more delayed bowling arm.
The intended endgame is stated: “The recommendations made by this model will be used to shape the future coaching of this individual. The results will be analysed and if positive the model will be developed into a coaching tool.” Note: no follow-up paper in this cluster reports whether that intervention was carried out or whether it worked.
What a coach should look for on video
Four cues, all in the front foot contact phase. This paper is the cleanest source for the priority ordering between them.
Cue 1 — At front foot contact: bowling arm back, front arm still up
- The cue: Position of both arms in the frame the front foot lands.
- Camera view + frame: Side-on, at front foot contact.
- What “good” looks like: Bowling arm delayed (extension not yet started); front arm still flexed/high, extension also delayed.
- What the fault looks like: Either arm already unwinding at landing.
- Why it matters: This was the most marked difference in the +22% optimisation, and the paper argues it is causally upstream — adopting this pose is what “allowed” the front leg to stay straighter and the trunk to flex more. If a coach can change only one thing, this is the frame to change.
Cue 2 — Front knee held straight through the phase
- The cue: Front knee from landing to release.
- Camera view + frame: Side-on, scrub front foot contact → ball release.
- What “good” looks like: Straight or straightening; hips and head not sinking.
- What the fault looks like: Knee flexing after landing.
- Why it matters: Present in every optimisation. Mechanically it is the brake that converts run-up momentum into rotation about the front foot.
Cue 3 — Trunk flexion increased
- The cue: How far the bowler folds forward over the front leg.
- Camera view + frame: Side-on, front foot contact → ball release.
- What “good” looks like: More fold than the bowler’s habit.
- What the fault looks like: Remaining upright, or folding sideways.
- Why it matters: Present in every optimisation; the model treats it as the consequence of the straight leg and delayed arm.
Cue 4 — The delivery is largely decided before it starts
- The cue: Rather than coaching the 0.1 s after landing directly, coach the shape the bowler arrives in.
- Camera view + frame: Side-on, at front foot contact.
- What “good” looks like: See Cue 1 plus a straight front knee already at landing.
- What the fault looks like: A good-looking landing pose is missing and the bowler is visibly “rescuing” the delivery afterwards.
- Why it matters: The +22% (landing pose changed) versus +10% (landing pose fixed) gap is the argument. The thesis states the same thing more strongly: in the optimal solution most muscle activations were constant, meaning performance was “pre-determined by the orientation of the body at front foot contact”.
Cue supported against an intervention: if a bowler at elite level is already inside a good S&C programme, this paper says extra strength is worth roughly a tenth of what technique is worth. Do not sell strength work as the primary speed lever for such a bowler.
Caveats and limits
- n = 1. One elite male fast bowler: age 18 years, mass 85.0 kg, height 1.94 m, England U19 team, identified as a potential England player. Same bowler as the thesis, the 2015 conference abstract and the 2020 journal paper. (Height is given as 1.94 m here and 1.935 m in the thesis — the same person, rounded differently.)
- The paper’s own framing is the caveat: it argues explicitly that group findings do not transfer to individuals. By its own logic, its numbers do not transfer to your bowler either. The pattern (straight leg, delayed arms, more trunk flexion) is what generalises; the magnitudes are one athlete’s.
- Simulation, not intervention. The paper promises to test the recommendations on the real bowler; no publication in this cluster reports that test.
- Planar (2D) model. Non-planar rotation approximated with massless segments.
- Front foot contact phase only. Whether the optimal landing pose is attainable — and what it would cost in the run-up and back foot contact phase — is outside the model.
- Conference paper. No confidence intervals, no per-component match table, no ground reaction force results reported.
Relationship to other Felton work
- Reports the same three optimisations on the same bowler as Felton 2015 — optimising fast bowling performance, with fuller method detail and the same 10% / 22% / 1% numbers. Both are conference presentations of PhD Chapter 9 (Felton 2015 — PhD thesis: factors limiting fast bowling, unrounded 9.8% / 21.5% / 1.3%).
- The peer-reviewed journal publication of this work is Felton 2020 — optimising the front foot contact phase.
- Cites Felton & King (2016) for the non-planar pelvis and trunk modelling method.
CONTRADICTION: (the individual-specific message is later partly reversed) This paper is the strongest statement of the individual-specific position in the cluster — group research “is not suitable to understand the changes required to optimise an individual’s performance”, and the output is a bespoke prescription for one named athlete. Six years later, 2023 commonalities work runs the same individual-specific method on ten elite bowlers and concludes it “has resolved the controversy on whether individual and group optimisation studies of fast bowling reflect underlying commonalities” — finding that the same handful of characteristics (more extended front knee, more flexed front and bowling shoulders at landing, delayed trunk flexion / shoulder extension / wrist flexion) emerged for every bowler. That is a substantial softening: optimal technique turns out to be far less individual than the 2017 framing implies. The 2023 paper does retain the caveat that whether an individual can adopt those characteristics depends on their own constraints — so the individuality survives in attainability, not in what the target is.
CONTRADICTION: (conference vs journal on the 22%) the 22% landing-position gain reported here does not appear in the peer-reviewed 2020 journal paper on the same bowler, which reports only 9.8% and states that varying the initial bowling arm position was outside its scope. The group equivalent, published in 2023, was 13.5%.